sign in sign up
<<
Home , Britannia, Lydney Park, Roman temple, Sacellum, Temple of Janus (Autun)

Shrines (Roman and Post-Roman) Introductions to Heritage Assets Summary

Historic ’s Introductions to Heritage Assets (IHAs) are accessible, authoritative, illustrated summaries of what we know about specific types of archaeological site, building, landscape or marine asset. Typically they deal with subjects which have previously lacked such a published summary, either because the literature is dauntingly voluminous, or alternatively where little has been written. Most often it is the latter, and many IHAs bring understanding of site or building types which are neglected or little understood.

This IHA provides an introduction to Roman and post-Roman shrines. Descriptions of the asset type and its development as well as its associations and a brief chronology are included. A list of in-depth sources on the topic is suggested for further reading.

This document has been prepared by Kate Wilson and Pete Wilson and edited by Joe Flatman and Pete Herring. It is one of a series of 41 documents. This edition published by Historic England October 2018. All images © Historic England unless otherwise stated.

Please refer to this document as: Historic England 2018 Shrines (Roman and Post-Roman): Introductions to Heritage Assets. Swindon. Historic England.

HistoricEngland.org.uk/listing/selection-criteria/scheduling-selection/ihas-archaeology/

Front cover The Carrawburgh mithraeum, Northumberland. Beyond the narthex (entrance area) the two areas of grass mark the position of benches with altar at the rear of the building. Between Chester’s and Housesteads Roman forts, it is an English Heritage property open to the public. Introduction

Roman-period shrines and are relatively common in comparison with those for the preceding and succeeding periods, largely by virtue of the more substantial nature of many of the remains.

Many aspects of Roman-period religion are of being for most in the Roman period and how well known, including its generally polytheistic religion cannot be readily separated from daily nature and its toleration and absorption of life. The diversity of Roman-period religion can most local religious practices and new faiths – perhaps be best seen through the variety of the alleged Roman aversion to Druidism may cults originating outside the Empire, or within have as much to do with the perceived political conquered territories, that are recognisable in influence of the as with Roman revulsion the archaeological record; major exotic deities in respect of human . Similarly the lack include Isis and Serapis from Egypt, of toleration of monotheistic faiths, notably and Atys from Asia Minor, and Mithras from the Judaism and Christianity, may have been as much . political, given the refusal of their adherents to acknowledge state cults and the acceptance of the Much of the evidence for religion in the Roman Roman order that such observance implied, as it period is in the form of portable material - was grounded in religious dogma. altars and explicitly religious artefacts such as statuettes of deities – as well as items that have The key Roman observances were the Imperial clear votive or talismanic significance, such as Cult, attested in Britain in various cities including phallic symbols. , and York, and the (, and ) who are In the post-Roman period shrines were used as well-evidenced. That said, most other major places of Pagan or (from the time of Augustine’s Roman gods are also represented in the British mission to Kent in 597) Christian worship until archaeological record, including , , the 7th century. The archaeological evidence for and . The equation of British such shrines is rare, with the majority of examples gods with gods from the Roman pantheon was so far identified being from the south-west of common, such as (the Celtic name of the England in rural, isolated settings, frequently goddess) and Minerva (the Roman equivalent located on hilltops. They appear as foundations thereof) at Bath, but by no means obligatory of small circular or rectangular buildings or as witnessed by the shrine to , an structures, usually found in association with other ‘un-equated’ Celtic goddess, at Carrawburgh features such as prehistoric sites, hillforts, Roman (Northumberland). temples, cemeteries and churches, and natural features such as springs and caves. The multiplicity of gods and spirits (such as the Genii loci – Spirits of Place, including the Lar Although sites may appear as cropmarks in aerial familiaris who presided over the household) photography or as low earthworks, their identity demonstrate how religion pervaded all aspects would need to be confirmed by archaeological

1 excavation. Very few sites have been identified of Anglo-Saxon words for , , through archaeological investigations in England idol, holy place or shrine as found in Harrow and of those only Yeavering (Northumberland) (Middlesex) and Weeford (Staffordshire). Place- provides evidence for pagan worship. Place- names incorporating these pagan elements are name evidence can also be used to identify concentrated in the midlands and south-east potential sites that may be indicative of pagan of England. places of worship because they contain elements

2 1 Description

For the Roman period it is perhaps best to separate the description of temples from shrines, the latter taken here to mean slighter or less formalised religious structures. There are three basic plan forms of temple: Romano-Celtic, Classical and Basilican, although as will be discussed below these generalised categories are subject to many exceptions. Shrines are far more varied in form, ranging from formally constructed buildings such as the semi-circular structure associated with the temple at Coleshill, (Warwickshire), through minor structures such as Figure 1 nymphaea, to other foci, for example the Bronze West Hill, Uley, . Extract from the site phase plans showing some developments within the Age barrow at Irlingborough (Northamptonshire) shrines complex. that became a focus for coin offerings, or the Iron Age Ferry Fryston (West Yorkshire) ‘chariot burial’ barrow which attracted offerings of cattle for up to 500 years, into the 2nd century AD. While 28 m square. On balance polygonal-plan temples perhaps not technically ‘shrines’ the house church tend to be larger than square-plan examples. at Lullingstone (Kent) Roman villa and other religious structures associated with Christianity Romano-Celtic temples occur in both urban and represent a further group of Roman-period rural Roman-period locations. Some form part of religious structures. religious complexes where there appears to be continuity of religious focus from the Iron Age, as Romano-Celtic temples represent by far the most at Hayling Island (Hampshire.) where a circular common plan form. In essence the main structure Iron Age ‘shrine’ was succeeded in the Roman often consists of two concentric elements, with period by a circular stone-built structure with a the inner one forming the or religious in place of the shrine. At the Springhead focus, and the outer square an ambulatory temple complex in Kent there appears to be around it. In plan they can be square, circular, no clear evidence of structural continuity from rectangular or polygonal and the most common the Iron Age to Roman period, but nevertheless reconstruction is as a central tower surrounded the springs on the site seem to have been by a lower ambulatory (Figure 1). They often, but venerated throughout. not without exception, stand within a temenos, or sacred enclosure. Considerable structural variety in Romano-Celtic temple forms have been observed. At Hayling In size they vary considerably, with the outer Island the lacks an ambulatory dimensions of ambulatories varying from about immediately around the cella, but has rather a, 8.5 m to about 22 m and those of cellae from probably internally porticoed, enclosure 40 m about 5.1 m to about 16 m. Continental examples by 43 m. At Hill, near Weymouth (Figure can be larger, for example the ‘Temple of ’ 2), the temple similarly lacked an ambulatory, at , in , the ambulatory of which was although here this probably reflected the

3 Figure 2 Jordan Hill, Dorset. The surviving remains of the cella of the 4th century AD Roman temple. Near Weymouth, this English Heritage property is open to the public.

robbing of the site. Equally structures can building height, 3.5 m at Colchester, to create a display distinctive additional features, such as visual impression of dominance of the area as the external buttresses known at Pagan’s Hill, well as supporting the cella. The latter was usually Somerset, while broadly conforming to the usual located behind a portico, or pronaos, that was plan forms. often approached by steps. The sides of the cella would usually be adorned with engaged or free Classical temples are rectangular stone-built stone columns, and the frontage would include structures that are relatively less common than a . Romano-Celtic temples, but are an important element in the religious landscape of Roman The extent of decoration is variable, although Britain. Of iconic status is the temple of the columns of both Corinthian and Iconic orders are deified at Colchester that is today known to be associated with Classical temples represented by its 32 m by 23.5 m podium which and there could be much use of marble and stone forms the platform for the Norman keep of panelling. The outer long sides of the podium Colchester castle. could be lined with free columns supporting the roof, following the plan forms well known While little more is known about the plan of the from the Continent or, as in the case of Wroxeter Colchester temple, that to Sulis Minerva at Bath I (Shropshire), appear rather as a compromise is better understood. The podium of the structure between the classical and Romano-Celtic plans is in essence an infilled vault that gave the whole incorporating a porticoed cella and entrance,

4 both probably with free columns and , The inclusion of mithraea in the group but separated by an elongated internal courtyard emphasises again the lack of homogeneity of the with porticoes on two sides reminiscent of type – the Mithraic rite demanded a structure of Hayling Island. a particular form, with a floor below ground level simulating a cave, benches along the sides for The temple would usually be located within devotees and a focus on the cult of Mithras a temenos, or sacred enclosure, although it is killing the bull. Other cults were less distinctive, possible to argue that the whole of the walled and recognition of the structure by form alone is area of Bath was a religious complex. The less clear. However, there is every likelihood that identification of ‘Temples’ I-VII at Corbridge has votive deposits (items left for purposes) been challenged by various authors, although are associated with Romano-British temples architectural fragments do suggest the presence and the identification of depositional practices of classical temples somewhere within the town. could help with the recognition of temple sites. Cults, such as that worshipped in the 4th century Basilican temples are conventionally oblong temple at Silchester, may have demanded light apsidal structures that represent a functionally and the inclusion of a clerestory in contrast with defined sub-group of a standard Roman-period the mithraea. building form – similar plans can be observed in military headquarters building (principia) and Known basilican temples are located largely in in towns. the south and close to the northern frontier, but given their basic plan form and relative lack The key element of a basilican temple is the of distinctive characteristics examples could presence of an aisled nave and many have a be going unrecognised. Where known, they square or semi-circular ‘apse’, and some a narthex are generally constructed of stone at least at or antechamber between the entrance and nave. foundation and lower wall levels, although given The size can vary considerably from the London the limitations of the evidence the upper parts mithraeum at about 20 m by 8 m to around 8.5 of the walls could have been half-timbered. m by 6 m for the first phase at Carrawburgh Internally, timber could have been used in aisle (Northumberland; Figure 3). arcades, at least in smaller examples and was certainly used in some cases for screens between narthex and nave as well as for flooring and benches in mithraea, although other flooring materials including stone slabs, tessellated pavements and mosaics are also known.

Other temples are in many cases essentially variations on the three main types, but sufficiently different to not readily fit the basic models. The temple to Antenociticus at Benwell (Tyne & Wear), 7 m by 5 m with a small semi-circular apse, is essentially basilican in plan form, but the scale precludes regarding it as such. Equally the Triangular Temple at Verulamium (St Albans, Hertfordshire) represents a pragmatic response to Figure 3 the limitations of its site, incorporating a possible The Carrawburgh mithraeum, Northumberland. Beyond pedimented entrance giving on to a possibly the narthex (entrance area) the two areas of grass porticoed courtyard leading to a cella or shrines mark the position of benches with altar at the rear of the building. Between Chester’s and Housesteads at the rear of the complex – all elements that can Roman forts, it is an English Heritage property open to be seen elsewhere, but not in a comparable form. the public.

5 The temple from the northern suburb of Catterick At Westhawk Farm, Ashford, Kent, excavation (North Yorkshire) can be seen as either adapting revealed a rectilinear ditched enclosure a Romano-Celtic plan albeit with an elongated surrounding a polygonal post-built structure cella, or as a variation on a classical form, in interpreted as a shrine; that would probably either case probably lacking a temenos. not have been understood if it had not been totally stripped, revealing a probably associated Shrines is a widely interpreted term, and sites waterhole which had received votive deposits such as Nettleton (Wiltshire) where a substantial (that is items left for ritual purposes). The ritual octagonal building is generally described as use of features such as waterholes is documented a ‘shrine’ to , illustrate the problem of elsewhere, for example at Shiptonthorpe, in the definition. The central structure, while not readily East Riding of Yorkshire, and suggests that Roman fitting the conventional plan of any of the ‘temple’ period religious foci that lack readily identifiable categories above, is built on a scale that suggests structural features may escape recognition unless that it is in effect a variation of a Romano-Celtic deposit survival and serendipitous excavation temple. produces evidence of votive deposition.

Similar issues appear in a later Roman context In military contexts the headquarters building where significant religious elements have been incorporated the regimental shrine ( recognised on a number of what are apparently or aedes), while urban sites such as Cirencester villa sites. The unpublished excavations at have produced Jupiter Columns that would have Littlecote, Wiltshire, illustrate this – what appears stood outside in the street or at another public to be a relatively modest villa on the Silchester location and which appear to have had a public to Mildenhall road had its baths enlarged in the religious function. mid-4th century and a ‘triconch’ hall constructed and furnished with an elaborate Orpheus mosaic. Christian churches could also in essence be It has been suggested that this may have been viewed as shrines with a particular dedication, a focus for the cult of Bacchus, although it may and they often conform to plan forms as be no more than an elaborate summer dining described above. For example the probable room, illustrating the extent to which religious Silchester church is a west-facing basilican iconography and possibly religious practice structure incorporating a narthex and semi- infiltrated daily life. circular apse, but is marked out by virtue of the presence of a baptistery outside the narthex and The shrine at the Coleshill Temple has already widened transepts at the west end of the aisles. been mentioned and secondary shrines are well known on temple sites, such as that dedicated Other structures are less visually distinct, such to the local nymphs and outside the as the possible church at Icklingham, Suffolk, mithraeum at Carrawburgh (Figure 3). a simple rectangular building associated with a possibly Christian cemetery, a suggested Rural shrines are a difficult area – the well- baptistery and two baptismal tanks. Various known site on Scargill Moor, near Bowes, County rectilinear structures at pagan temple sites have Durham, consists of two isolated shrines with been suggested as later Christian churches, such altars dedicated to Vinotonus, who on one altar as at Uley, Gloucestershire. is conflated with , perhaps suggesting an association with hunting. Without the altars any On military sites churches have been suggested recognition of this site as religious would have at Richborough, Kent, there associated with been difficult, and it is possible that small rural a baptistery, and at various sites on or near shrines could take many forms. ’s Wall, including apsidal structures at South Shields, Housesteads and . However, the best attested church in Britain is

6 not a separate structure, but rather the house- millennia from the to the 7th century AD. church at Lullingstone evidenced in large part by Others were built as completely new constructions fallen wall plaster featuring chi-rhos and images in the 5th or 6th century AD and were used for of probable Christian worshipers, rather than only a short period of time. An example is an distinctive structural evidence. Other house- early 7th century AD building excavated at the churches may be indicated by the presence of Northumbrian palace site of Yeavering. This was Christian imagery, such as the probable head rectangular, about 12 m x 6 m internally, and was of Christ from Hinton St Mary, Dorset, while at interpreted as a temple because of its association the Frampton villa, also in Dorset, structural with a pit full of ox skulls, and human graves, components of the site recall a basilican plan with clustered to its south (Figure 4). an apse and narthex and the decoration includes a chi-rho.

Sites of Uncertain Function There were clearly religious monument types that no longer survive, such as the beehive structure known to have survived until 1743 – Arthur’s O’on at Carron on , which may have been a temple to Victory, or given its location north of The , a tropaeum, an official monument dedicated to Victory.

The temple mausoleum at Lullingstone may indicate another area of uncertainty with Figure 4 commemoration of the dead being aggregated Yeavering, Northumberland. Possible 7th century shrine. with religious observance, in which case it may be simplistic to take as read the primary burial function of all features claimed as mausolea, Once shrines went out of use they appear to particular given their potential similarity to have been abandoned or incorporated into later Arthur’s O’on. ecclesiastical buildings. Because of their close associations they can sometimes be confused Post-Roman shrines are small circular or with earlier shrines and temples as well as the rectangular wooden or stone structures or remains of small structures of a later date with buildings which provided the setting for Christian quite different functions. However, they can be or Pagan worship. Post-Roman shrines are very distinguished by careful analysis of associated rare and at present we know of few examples. artefacts and stratigraphical analysis resulting These probably represent only a fraction of the from archaeological excavation. shrines which existed and which are frequently referred to in contemporary literature by Bede We still know little about the landscape beyond and others; many have probably left little the shrines themselves and what links they may archaeological trace or the evidence lies under have had with contemporary settlements and Anglo-Saxon or later churches and monasteries. society beyond an assumed local or individual responsibility for their maintenance. There are The main components of a post-Roman shrine many literary references to English pagan worship are building foundations or foundation trenches, in, for example, Bede’s Ecclesiastical History of graves, free-standing posts or platforms which are about AD 731. These frequently refer to altars and sometimes set within an enclosure wall or bank. idols and occasionally to heathen rites. However, Some shrines re-used earlier sites and buildings they rarely describe pagan sites or provide and show remarkable continuity in the use of topographical detail linking them to known the place or site for ritual purposes over four archaeological sites.

7 2 Chronology

In chronological terms many of the sites of development of new pagan shrines. The temple Roman-Celtic temples that have been subject to at Park, Gloucestershire, for to excavation can be traced back to the late instance, was founded in the late 3rd or early 4th Pre-Roman Iron Age, and on some sites, such as century and enjoyed considerable investment of Hayling Island, continuity from the pre-Roman resources in the later 4th century. Other pagan period has been demonstrated. The same may sites remained in use into the late 4th or 5th be true for the religious centre at Bath where an century, and while destruction by Christians Iron Age focus has been suggested to precede has been claimed at some the evidence is the classical temple to Sulis Minerva. On the often equivocal. other hand, although pre-Roman religious sites are known in the Colchester area, for example at Post-Roman shrines were in use from the 5th Gosbecks, the Colchester temple of Divus Claudius to the 7th centuries, after which Christianity was an entirely new creation designed to send a and worship in churches became widespread political message demonstrating the existence of throughout England. Individual shrines were in the new order. Construction was probably in hand use for varying lengths of time and very few have by AD 54 in Nero’s reign, and it seems that once been securely dated. Because of this there is little established major temples remained in use for evidence for any chronological development. much or all of the Roman period. Most examples have been identified during the archaeological investigation of associated Whilst the Edit of Milan in AD 313 allowed features such as hillforts, cemeteries, Roman for religious toleration, seemingly allowing temples and cathedrals, and these earlier and Christianity to flourish without the threat of later structures and features have provided persecution, it did not bring an end to the relative dating evidence.

8 3 Associations

In the Roman period temples and shrines are apparent isolation, or represent one element in found on most site types. In urban locations these a complex multi-functional townscape in a vicus can be extremely prominent at the heart of the associated with a fort. settlement, as well as alongside major roads. A temple could be the focus for a considerable Post-Roman shrines are associated with a variety complex, as at Nettleton, Wiltshire, and it can be of contemporary buildings and features including argued that the religious aspects of ‘small towns’ cemeteries, churches, settlements, roads, and such as Bath and Springhead, Kent were their trackways. They can also be associated with non- dominant feature, in the case of Springhead the contemporary features including Roman temples, religious elements being represented by a group , hillforts, prehistoric burials and of at least seven temples. ritual sites and medieval churches, monasteries and cathedrals. They are frequently found in A major temple might be associated with any prominent positions on hilltops or close to Roman combination of: subsidiary shrines, a guesthouse, roads and trackways, and may be associated associated service structures, cemeteries and a with natural features such as springs, wells, rivers settlement to accommodate those serving the wetlands and caves. They suggest a wide array of temple. Alternatively a temple or shrine can sit in sacred natural places in a wider ritual landscape.

9 4 Further Reading

For the Roman period the key corpus of structural For the post-Roman period there is no complete evidence is M J T Lewis’s Temples in source book available but R Morris, Churches in (1966). the Landscape (1989) provides an overview of the evidence for the transition from Pagan to Christian Other important publications include: worship during this period, and A Woodward, Shrines and Sacrifice (1992) provides a general W Rodwell (ed), Temples, Churches and Religion: introduction to the topic. Recent Research in Roman Britain (1980); A more specific analysis of the evidence can M Henig, Religion in Roman Britain (1984); be found in J Blair, ‘Anglo-Saxon Pagan Shrines and their Prototypes’, Anglo-Saxon Studies in and for Christian sites D Petts, Christianity in Archaeology and History 8 (1995), 1-28, and for Roman Britain (2003). the south-west P A Rahtz and L Watts, ‘The End of Roman Temples in the West of Britain’ 183-210 in Important site-specific studies include: P J Casey (ed), The End of Roman Britain (1979).

R E M and T V Wheeler, Report on the Prehistoric, One site which provided detailed archaeological Roman and Post-Roman site in Lydney Park, evidence is reported in B Hope-Taylor, Yeavering Gloucestershire (1932); (1977).

P D C Brown, ‘The Church at Richborough’, For a brief summary of the place-name evidence 2 (1971), 225-31; D M Wilson, ‘A Note on OE hearg and weoh as Place Name Elements Representing Different B Cunliffe and P Davenport, The Temple of Sulis Tof Pagan Worship Sites’, Anglo-Saxon Studies in Minerva at Bath (1985); Archaeology and History 4 (1985), 179-83.

G W Meates, The Roman Villa at Lullingstone, Kent. A more detailed study can be found in M Gelling, Volume II: The Wall Paintings and Finds (1987); Place-names in the Landscape (1984).

A Woodward and P Leach, The Uley Shrines (1993); See S Semple, ‘In the Open Air’, in M Carver, A Sanmark and S Semple (eds), Signals of Belief in J D Shepherd , The Temple of Mithras, London Early England: Anglo-Saxon Revisited (1998); (2010), 21-48, for a landscape-led approach to identifying ritual activity. and A C King and G Soffe ‘Hayling Island: a Gallo- Roman temple in Britain’, in D Rudling (ed), Ritual Landscapes of Roman South-East England (2007).

10 5 Where to Get Advice

If you would like to contact the Listing Team in one of our regional offices, please email: [email protected] noting the subject of your query, or call or write to the local team at:

North Region South Region 37 Tanner Row 4th Floor York Cannon Bridge House YO1 6WP 25 Dowgate Hill Tel: 01904 601948 London Fax: 01904 601999 EC4R 2YA Tel: 020 7973 3700 East Region Fax: 020 7973 3001 Brooklands 24 Brooklands Avenue West Region Cambridge CB2 8BU 29 Queen Square Tel: 01223 582749 Bristol Fax: 01223 582701 BS1 4ND Tel: 0117 975 1308 Fax: 0117 975 0701

11 We are the public body that helps people care for, enjoy and celebrate England’s spectacular historic environment.

Please contact [email protected] with any questions about this document.

HistoricEngland.org.uk

If you would like this document in a different format, please contact our customer services department on:

Tel: 0370 333 0607 Email: [email protected]

All information and weblinks accurate at the time of publication.

Please consider the environment before printing this document

HEAG 234 Publication date: v1.0 May 2011 © English Heritage Reissue date v1.1 October 2018 © Historic England Design: Historic England and APS.